The THY-MOD Study - Personalized Dosing in Children with Hyper- or Hypothyroidism Computed by Mathematical Modeling

甲状腺激素对于胎儿和婴儿的正常大脑发育，以及儿童和青少年的正常认知功能、正常生长和青春期至关重要。婴儿和儿童甲状腺疾病（先天性和获得性甲状腺亢进或甲状腺功能减退）的治疗是困难的。首先，有必要根据临床经验仔细选择初始剂量。其次，在甲状腺激素达到生理平衡后，需要在青春期和发育期根据年龄和甲状腺功能不断调整个体剂量，以维持甲状腺功能正常。为了减轻认知障碍等神经系统和发育不良后果的风险，必须建立一个最佳的、个性化的剂量，并对新生儿、婴儿和甲状腺功能减退或亢进儿童进行持续微调。基于个体协变量效应的群体药代动力学/药效学（PKPD）原理和最优控制理论的模型预测控制算法的动态数学模型将被开发并应用于来自不同医院的回顾性纵向疾病测量。开发的模型和算法将在本项目期间根据预期收集的数据进行验证。这将允许预测一个最佳的个性化给药策略，维持甲状腺激素在一个适当的参考范围。一般来说，计算婴儿和儿童最佳个性化剂量的模型和算法是优化长期神经学结果的医学必要性，从而提高学校和专业教育的表现，并提高儿科患者及其父母的生活质量。开发的最优控制算法将免费提供，用于教育目的，并可能应用于其他疾病领域。

Thyroid hormones are essential for normal brain development in the foetus and the infant, and for normal cognitive functions, normal growth and puberty in children and adolescents. Treatment of thyroid diseases (congenital and acquired hyper- or hypothyroidism) is difficult in infants and children. First, a carefully selected initial dose based on clinical experience is necessary. Second, after reaching a physiological balance of thyroid hormones, a continuous adjustment of the individual dose depending on age and thyroid function during adolescence and puberty is required to maintain euthyroidism. To mitigate the risk of negative neurological and developmental outcome such as cognitive impairment, it is essential to establish an optimal, personalized dosing that is continuously fine-tuned in neonates, infants and children with hypo- or hyperthyroidism. Dynamical mathematical models based on population pharmacokinetic / pharmacodynamic (PKPD) principles with individual covariate effects and model predictive control algorithms from optimal control theory will be developed and applied to retrospective longitudinal disease measurements from different hospitals. Developed models and algorithms will be validated based on prospective collected data during the duration of this project. This will allow the prediction of an optimal personalized dosing strategy that maintains thyroid hormones in an appropriate reference range. In general, models and algorithms to compute optimal personalized dosing in infants and children are a medical necessity to optimize long-term neurological outcome and consequently enhance performance in school and professional education, and to increase life quality of paediatric patients and their parents. Developed optimal control algorithms will be made freely available for educational purposes and for possible application to other disease areas.